A frequent problem encountered by resists used to process supports comprising semiconductor devices, is reflectivity back into the resist of the activating radiation by the substrate, especially those containing highly reflecting topographies. Such reflectivity tends to cause standing wave ripples and reflective notches, which interfere with the desired pattern in the photoresist. The notches are particularly bad when the support is non-planar. The conventional approach to this problem is to incorporate an anti-reflective dye either in the photosensitive layer or an adjacent layer.
Although there are many dyes known to be anti-reflective from their use as, e.g., antihalation dyes in the silver halide photography art, it is not possible to transfer such dyes wholesale to the photoresist arena. There are conditions peculiar to the photoresist art which render many such dyes unsuitable. One of these is the high temperature baking that is given to the resist, prior to its exposure. For example, it is common to force either the photoresist layer or the antireflective layer into a planar configuration by baking it until it melts and flows, creating a condition known as "planarizing." Such baking can require temperatures as high as 200.degree. C. Unless care is taken in the selection of the dyes, such high temperatures are deleterious in at least two respects: First, because the antireflective layer is very thin, e.g., no greater than about 5 .mu.m, the dyes tend to volatilize out of the layer, particularly if they are not so large in molecular weight as to be mordanted. Second, some of the dyes previously used in photography for antihalation are thermally unstable--that is, they bleach out when heated. Such bleaching is unsuitable when used with a resist, since the baking step described above would destroy the antireflective property before it can be used during the exposure step. An example of such bleachable dyes is the following dye, described in EPA 119,831: ##STR2## The mechanism of the bleaching of this dye is not explained, but it can be postulated to be similar to the electrocyclic ring closure of dyes of the type ##STR3## the ring closure mechanism of which is based on the presence of hydrogens in the mono-N-substitution, as described in J. Amer. Chem. Soc., Vol. 92, p. 5641-47 (1970).